As I understand your situation (correct me if I am wrong):

1. You want to show that you were able to reliably fractionate (and enrich for) each cellular compartment.

2. You want to show that you loaded an equal amount of the membrane fraction to your SDS-PAGE for each sample.

These are two different aims that may require two different solutions, though a single solution is possible.

For #1: use an antibody against a protein that is exclusively present on the membrane. A suggestion has been to look at cadherins, but other ubiquitous receptors would also suffice (gp130, some integrins, etc). Alternatively, choose a membrane protein that has been used in similar studies. Do not use actin. This is a cytoskeletal protein found intracellularly, although some complexes are formed with membrane proteins, using actin will not provide evidence for #1.

For #2: if you have shown #1 above convincingly with a membrane protein, then commassie blue would be okay to use here. However, better to use a second protein or the one used for #1 as long as you know that the protein levels are not affected by your treatments. Do not use actin. Again, you are trying to convince the reviewer that you have in fact achieved fractional enrichment, why then would you look at an internal protein?

Not knowing all the details of your experiment, I would recommend that you use gp130 for both #1 and #2. You'll have to determine if this is appropriate for your experimental model.

Best of luck.

It is better to use at least two housekeeping genes for internalcontrol. For plasma membrane, actin, Calnexin, cadherin, plasma membrane calcium ATPase (PMCA), and b-catenin are popular to be used for internal control of plasma membrane.

I have seen people use beta actin as a loading control and published in Oncogene. See: HPV-16 E5 oncoprotein upregulates lipid raft components caveolin-1 and ganglioside GM1 at the plasma membrane of cervical cells. F A Suprynowicz, G L Disbrow, E Krawczyk, V Simic, K Lantzky and R Schlegel

Oncogene 27: 1071-1078; advance online publication, August 20, 2007; doi:10.1038/sj.onc.1210725. Figure 5

I agree that it's helpful to use more than one internal control. Housekeeping protein levels are unfortunately not as stable we once thought. For whatever internal controls you choose, I recommend that you run an experiment to make sure that the levels of the control proteins are not affected by your specific experimental treatments and conditions.

It's important to check it yourself in your system, even if the literature says it's an appropriate control. Many studies have found poor correlation between mRNA and protein levels, so you really can't assume that stable gene expression will lead to stable protein levels in your hands. 

As I understand your situation (correct me if I am wrong):

1. You want to show that you were able to reliably fractionate (and enrich for) each cellular compartment.

2. You want to show that you loaded an equal amount of the membrane fraction to your SDS-PAGE for each sample.

These are two different aims that may require two different solutions, though a single solution is possible.

For #1: use an antibody against a protein that is exclusively present on the membrane. A suggestion has been to look at cadherins, but other ubiquitous receptors would also suffice (gp130, some integrins, etc). Alternatively, choose a membrane protein that has been used in similar studies. Do not use actin. This is a cytoskeletal protein found intracellularly, although some complexes are formed with membrane proteins, using actin will not provide evidence for #1.

For #2: if you have shown #1 above convincingly with a membrane protein, then commassie blue would be okay to use here. However, better to use a second protein or the one used for #1 as long as you know that the protein levels are not affected by your treatments. Do not use actin. Again, you are trying to convince the reviewer that you have in fact achieved fractional enrichment, why then would you look at an internal protein?

Not knowing all the details of your experiment, I would recommend that you use gp130 for both #1 and #2. You'll have to determine if this is appropriate for your experimental model.

Best of luck.

@Jiaqi Shi

In the reference you provided the investigators use actin for what appears to be whole cell lysates (Fig 5a). For their fractionation blots they simply state that "Equal amounts of total cellular protein were loaded."

@Nirmalya

If your experiments are similar to what Jiaqi referenced then I don't see why commasie blue would not be appropriate. If possible, provide a figure of your results so that we can provide a more specific and accurate answer.

Thank you all for your support.

@ Ricardo

Thanks for your response. I showed the up-regulation of caveolin-1 during some treatment. When I used whole cell lysate, the signal  is very poor in WB. Hence I use plasma membrane fraction to enrich the caveolin-1.  I am detecting the membrane caveolar protein caveolin-1, I think it will justify your point #1. However, my only aim was enrichment of caveolin-1 for getting a stronger signal. 

For point#2: Now, the problem is that caveolin-1 is interacting with large number of proteins in membrane like the common markers of membrane such as Cadherins, catenins, Na-K-ATpase. Up-regulation of Caveolin-1upon treatment  may pull more of these interacting protein in membrane which leads a confusing data after normalization. For this reason, I used the PVDF coomassie stain.

Finding new proteins in membrane which can act as housekeeping for my system  is difficult for me in this stage.  For instance,   I know from my own data that  integrin b1 expression is  altered in my system. Still I will go through the literature what you and others suggested me to use as housekeeping.

I am absolutely agreed with your opinion for  using b-Actin, although some of my colleagues insisted to use actin. Can you give me any reference which states that Actin can not be use in plasma membrane or it is a contamination in plasma membrane.

If you provide me your mail ID, I will send you the figures asked by reviewer for correction. My ID is [email protected].

Thanks again,

Regards

Strictly speaking, actin isn't necessarily a contaminating protein in your membrane fraction. The presence of actin is simply due to the nature of the method (differential centrifugation). Doing a WB of actin would more likely show that your samples have different levels of cytoplasmic "contamination", but would not in itself show that you loaded equal membrane protein amounts. By using actin you would be assuming that the level of "contamination" is the same for every sample, and I don't think this is a fair assumption. This is probably why the investigators did not use actin for their fractions in the study that Jiaqi provided.

You could add your figure to your original post by editing it. This way we can all comment on it. Alternatively, please send me a PM through ResearchGate, as this will allow me to keep my ResearchGate e-mails organized.

Once, for a similar experiment, I used a kit from Thermo Scientific and they suggested me to use the EGFR as a marker for the membrane. EGFR more or less is expressed by all the kinds of cells but obviously its expression really depends on the sort of treatment you are doing.

Transferrin receptor is a good control for transmembrane protein.

good luck!

NaK-ATPase (alpha-subunit)

Pan-Cadherin

 Hi Nirmalya, what protocol did you use for the isolation of the membrane fraction? 

We are trying to isolate plasma membranes from the frozen muscle tissue (rat uterus). Any suggestions?

Thank you! 

Nirmalya, you may want to try total protein staining as your internal control rather than a housekeeping gene.

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0072457

@Amy

I used it  in my initial submission. But reviewers raised objection on total protein as loading control. Then I used membrane bound actin as loading control (it may come as contamination during isolation process). Fortunately, I got the equal band of actin which satisfied the reviewers. However, I had a plan B, if I didn't get that I will justify my total loading control citing several papers including this paper also.

Interesting that the reviewers objected to the total protein loading control. Could you send me a link to your paper? I've been digging into the WB normalization issue recently (planning to write a review) and it would be really helpful to see the context of your experiments. Thanks!

@Amy

The Pubmed link of my paper-

http://www.ncbi.nlm.nih.gov/pubmed/?term=25842166

If you want I can send you the earlier figure also, where objection was raised.

I ahve stumbled upon similar problem. In my membrane prep I see equal amount of beta actin. But planning to probe for other membrane resident protien to be safe.

More evidence that housekeeping proteins may not be the most reliable choice for Western blot normalization -

Article Use of the REVERT ® total protein stain as a loading control...

Bio-Rad Stain free Gel can be a good option. It clearly shows the loading and running status after running and transfer. Both Gel and Membrane can be used.

Normalization in Western Blotting by a house-keeping protein is a completely erroneous technique, but liked by the reviewers in general. It is possibly meaningful if both the target protein and the loading control produce signals in the linear range. In most cases , however, the house keeping protein being highly expressed, shows saturation effect and is not really acting as a loading or transfer control. And also Western blot signals may not follow any linear relationship with the amount of protein. In ELISA, in many cases we do not find any linear range in calibration curves which means that n-fold increase in signal intensity may actually mean much more or much less than n-fold increase in protein. Western blotting is based on similar principle as in ELISA. So what normalization are we talking of if target protein and loading control have different types of calibration curves and they are in different regions of the calibration curves ? Normalization by total protein staining is a much better option.

It is ridiculous that if we quantify a protein by Western blotting, the reviewers ask for a loading control. If we quantify the same protein by ELISA, they don't ask for anything. In one case we load in a gel-slot and in another in a plastic-well.